Re: compiling/builting 2.1b1 on hpux 9.7 and others

[Kannan Varadhan <kannan@ISI.EDU>]

Jon,

>>> From: Jon Crowcroft <J.Crowcroft@cs.ucl.ac.uk>
>>> Date: Fri, 23 Jan 1998 16:21:01 GMT

> some other things - we have 3 bits of ns work (real work) i guess we'd
> like on the web page -
> rmdp/rlc work lorenzo vicisano did see
> ftp://cs.ucl.ac.uk/darpa/infocom98.ps
> multcp work philippe oechslin did,see
> ftp://cs.ucl.ac.uk/darpa/multcp.ps
> 
> and isidor's work on self organising transcoders.....reference to follow
> soon. But what is the preferred way for me to submit this (e.g. the
> .cc/tcl and script files etc?
> )

For the refs, please send us a citation, possibly category
that you'd like it classified under, and we can put it up on
the research using ns pages,

Here's what we suggest for code...maybe we should put this up on 
the web pages...


Kannan

----------------------------------------------------------------
To: Christoph Haenle <chris@cs.vu.nl>
Cc: ns-users@mash.CS.Berkeley.EDU
Subject: Re: MFTP port to ns-2 
Date: Tue, 09 Dec 97 09:37:23 PST
From: Kannan Varadhan <kannan@owe.isi.edu>

Hi Christoph,

It is great to have other people's developing code into the core of ns.
There are two modes in which we can use your work:

	a)  You put your sources on the web, and we add a pointer to it
	    on the ns-contributed code pages.  In this mode, you can 
	    provide the patches for items 1 through 4 directly for people
	    to use.

or alternately,

	b)  we can add it as part of the distribution.  You would be
	    nominally responsible for questions or bugs related to your
	    code.  If you have mods at a later date, you could send
	    us patches, and we will batch and apply them periodically.
	    In this case, you should send us a tar file with your sources,
	    as well as patches (source code diffs) to existing files.
	    A nominal README would be nice, as is an example of how to
	    use your modules in user scripts.  Further optional (but
	    useful) extras are validation tests/documentation of your code.

Thanks for your contribution,


Kannan (for the VINT/ns developers team)

PS:  As for the RM code, it looks like I'd be the stuckey responsible
for integrating into ns.  If so, I'd like to wait until mid-January.
I trust that will be ok with you? -- Thanks.


> Hello ns-2 developer team,
> 
> you said that you are glad about ports of existing multicast protocols to ns-
    2.
> I've ported the core of MFTP's file transfer phase ("Multicast File Transfer 
> Protocol"). Also implemented is the erasure correction enhancement that 
> I have been talking of at the IRTF-meeting in Cannes. Don't think my sources 
> are stable enough for a permanent integration it into ns-2, but they could 
> well yield an optional an extension package.
> 
> For the extension package, I'd need the following from your ns-developers:
> 
> (1) a payload-type "PT_MFTP" in file "trace.h", which could be permanently 
> assigned (i.e. even be defined if the extension package is not used).
> 
> (2) an entry "{ MFTP off_mftp_}" in the foreach-loop of "Simulator instproc 
> create_packetformat" in file ns-packet.tcl, which again could be installed 
> permanently assigned.
> 
> (3) 4 instructions:
>       source ../mcast/mftp.tcl
>       source ../mcast/mftp_snd.tcl
>       source ../mcast/mftp_rcv.tcl
>       source ../mcast/mftp_rcv_stat.tcl
>     in file ns-lib.tcl (which must only be added manually if the
>     MFTP-package is used).
> 
> (4) 4 entries in Makefile.in (OBJ_CC:)
>       mftp.o \
>       mftp_snd.o \
>       mftp_rcv.o \
>       codeword.o
>     Again, these four lines should not be present unless MFTP is used.
> 
> 
> There is further info (incl. demo script and tracefile) available under
> http://www.cs.vu.nl/~chris/Thesis/mftp/mftp_readme.html
> 
> 
> Do you have any suggestions on how to seamlessly integrate the extension pack
    age on demand into ns-2?
> 
>      -Chris.
----------------------------------------------------------------

--
kannan@isi.edu
http://www.isi.edu/~kannan